Replay Engine

Start Replay

Re-process all 6 engines (Feature → Gamma → Stress → Seller → Gap → Scoring) for a historical market session. Requires tick_data for the target date (run Backfill first if needed).

📊 Current Job

No replay jobs found

📋 Replay History

IDDateStatusProgressErrorsSignalsStarted (IST)Finished (IST)
Loading...

📖 How Replay Works

DAG Order: Feature → Gamma → Stress → Seller → Gap → Scoring

Per minute bucket (09:15–15:30 IST):

  1. Read tick_data & oi_data for the minute
  2. Compute features (delta, IV, OI, liquidity metrics)
  3. Run BSM gamma → gamma_summary
  4. Run BSM delta → stress_summary
  5. Aggregate seller stress → seller_stress_summary
  6. Compute gap analysis → gap_summary
  7. Score all signals → scoring_signals

Prerequisites: tick_data must exist for the date. Run Backfill to fill MQTT gaps first.

Idempotent: Re-running overwrites previous engine outputs (ON CONFLICT DO UPDATE).